The workhorse of AEO
Posts written for the question an engine was asked
A blog post is the most flexible unit of AEO content and the easiest one to waste. The post that earns a citation answers one buying query completely, in its first paragraph.
Every plan queries all 5. See how each engine picks who it names
The number behind this page
of the pages ChatGPT cites carry structured data, and 65% of the pages Google AI Mode cites do (SE Ranking, 2026). The cited pages are specific and organized rather than broad.
The gap this closes
- The blog already exists, is updated, and does nothing for vendor selection.
- It answers the questions you find interesting, not the ones buyers ask an engine.
- Buyers write full sentences carrying constraints: headcount, industry, compliance, the situation.
- A post can be correct and still unusable, with the answer buried in paragraph nine.
What you end up with
- Finished drafts aimed at named queries from your scan, not a content calendar.
- The answer committed in the first paragraph, which is the part an engine lifts.
- One buyer question fully answered per post, in your voice, ready to publish.
- Published straight to your CMS, then re-scanned so you can see if it moved.
How it works
Three steps.
The question comes from the record, not a brainstorm
Each draft is aimed at a buying query your scan already recorded as lost, in buying-intent order. The brief carries the exact phrasing the engines were asked, the competitors they named instead, and the verbatim answers. The draft is written against a real target, never a guess about one.
The draft commits early and stays checkable
The question becomes a heading. The answer lands in the first two sentences, and the rest of the page earns it with specifics a reader can verify. The FAQ block goes in with its markup, the internal links go in, and the draft arrives ready to paste rather than ready to brief.
You publish, and the next scan referees it
Publishing runs through your CMS connection, so the post lands on your domain under your byline. The next scan re-asks the query the post targeted on all five engines. That is the only honest test, which is why every draft targets named questions from the start.
What you get on each plan
You get finished drafts aimed at named queries, not a content calendar. This is what lands, per plan, with the monthly volume read live from the plan configuration.
Track
Free- No drafting on this tier, and we will not pretend otherwise
- Every question where a competitor was named instead
- Ranked by buying intent, competitor names attached
- The topic list a paid plan would draft, shown undrafted
- Next free scan re-asks the same queries
Fix
$99/mo- Losing queries drafted, up to your monthly cap
- Answer-first, with the FAQ block already marked up
- Internal link plan pointing at your money pages
- The engine answer that lost it sits alongside
- Publishes to your domain, next scan referees it
- Tracked from idea to published to cited
Dominate
$499/mo- Wider budget reaches further down your buying queries
- Higher draft volume against the questions you are losing
- Every brief reused by the video and newsletter generators
- One buyer question answered in three formats
Rule
$1999/mo- Dominate's numbers, with our team doing the work
- We review each draft before it reaches you
- For a firm with nobody in-house to publish
- Publishing to your site only if you grant it
Rule is portfolio level and sales-assisted, which is the right conversation if you are running this across a book of clients.
Which engine actually leans on this
Your row is Your own site. The rest of the grid is on the engines page.
| Channel | ChatGPT | Claude | Perplexity | Gemini | Google AI Mode |
|---|---|---|---|---|---|
| Your own site Schema, answer pages, llms.txt, crawlability | Heavy Vendor | Heavy Vendor | Heavy Measured | Heavy Our read | Heavy Vendor |
| Reviews and directories G2, Clutch, Capterra, Trustpilot, Google Business | Moderate Our read | Light Measured | Heavy Our read | Moderate Our read | Moderate Measured |
| Reddit and forums Unpaid opinion, where buyers go for the unvarnished view | Light Our read | Light Measured | Heavy Our read | Moderate Our read | Heavy Measured |
| Video YouTube. Only counts once there is a transcript | Moderate Our read | Light Our read | Moderate Our read | Heavy Our read | Heavy Measured |
| Newsletters LinkedIn, Substack, Beehiiv. Published under a named person | Moderate Our read | Light Our read | Moderate Our read | Heavy Our read | Heavy Measured |
| News and PR Earned coverage, wire pickup, journalist mentions | Heavy Vendor | Heavy Vendor | Heavy Our read | Moderate Our read | Moderate Our read |
Your own site. Your own site counts most for ChatGPT, Claude, Perplexity, Gemini, Google AI Mode.
Reviews and directories. Reviews and directories counts most for Perplexity. It is a light signal for Claude.
Reddit and forums. Reddit and forums counts most for Perplexity, Google AI Mode. It is a light signal for ChatGPT, Claude.
Video. Video counts most for Gemini, Google AI Mode. It is a light signal for Claude.
Newsletters. Newsletters counts most for Gemini, Google AI Mode. It is a light signal for Claude.
News and PR. News and PR counts most for ChatGPT, Claude, Perplexity.
Where this comes from
- OpenAI: publishers and developers FAQ States that a site must allow OAI-SearchBot to be eligible for ChatGPT search at all.
- OpenAI: introducing ChatGPT search The named publisher partnerships behind the news and PR row.
- Anthropic: web search tool documentation Citations are always enabled. Left to itself Claude decides when to search; our scans force the search on every query.
- Google Search Central: AI features and your website AI Overviews and AI Mode report inside the ordinary Search Console Performance report.
- Search Console Help: impressions, position and clicks How AI Mode counts a click, and why a follow-up question is treated as a new query.
- Dan Petrovic: citation mining of OpenAI grounding metadata Reddit retrieved in 76 percent of OpenAI searches and selected 0.61 percent of the time.
- Our production scan database, read 12 August 2026 46 completed reports, 34 brands, 419 buying queries, 2,421 answered buying cells. Where the per-engine mention rates come from.
- Our 15-region managed services study 743 records, 566 named by an engine, and the per-engine corroboration rates.
- What our crawler finds on real sites 39 readable crawls: 38 with no llms.txt, 26 with no FAQ schema, 16 with no Organization schema.
- How we check an AI visibility report The method behind every first-party number on this page, including what it cannot show.
Engine behaviour moves month to month. This table is maintained, not published once.
The proof, including the parts that flatter nobody
We push structure this hard because we can see how thin the average result is, on firms that arrived already suspecting they were missing.
Across the 46 completed reports behind this page, the best any engine managed was naming the brand in 7.4% of the buying queries we put to it, on those firms that suspected a gap.
The others sat between 0.4% and 4.0%, on the same firms that suspected they were missing.
Our July managed IT study says something sharper about why retrieval matters:
In July, Perplexity, which retrieves before it answers, produced 101 of 257 names we could resolve to a real firm. In the same managed IT study, Claude, then answering from memory, produced 2 of 129.
Every engine we query now searches the web first. A published page can reach all of them, and a page that does not exist reaches none.
What we do not claim: that a specific post caused a specific citation. We can show the query it targeted, the engines asked before and after, and whether the answer changed. That is correlation with a date stamp: more than a content report usually offers, and less than proof.
Read the MSP AI Visibility Report 2026 →A worked example: the brief a draft is written from
Nothing here is a template with blanks. A brief is a record pulled out of your scan. This is one, with an illustrative brand and an illustrative engine answer standing in for a real customer's.
Brief: content brief 3 of the month
| Field | Value |
|---|---|
| Target question | which IT provider should a 40-person accounting firm use for SOC 2 readiness |
| Engines asked | Every engine on your plan. Most answered, one returned nothing, one named you in passing |
| Verdict | Absent on ChatGPT, Claude and Google AI Mode; brief mention on Perplexity |
| Named instead | Vendor A on three engines, Vendor B on two, Vendor C on one |
| Quote from the losing answer | "Look for a provider with prior SOC 2 Type II experience in professional services and a named compliance lead." |
| Angle the quote hands you | Answer the criteria the engine itself stated, with evidence, before anyone else does |
| Funnel position | Buying intent, weighted highest in the Index |
What the draft is required to contain
- H1 phrased as the question, so the match is not left to inference.
- A committed answer inside the first two sentences, before any context.
- The five artefacts an auditor asks for, named, because the engine's own answer asked for evidence.
- A short table of what a 40-person firm needs versus a 400-person firm, since headcount is the qualifier in the question.
- An FAQ block of four pairs, marked up, matching the visible text word for word.
- Internal links to the SOC 2 service page and the comparison page naming Vendor A.
You see the brief before anything is written, and you can reject it. That matters more than it sounds. The brief is where a wrong assumption about your business is cheap to fix; a draft is where it is expensive.
How it actually works
The mechanism, for anyone who wants it. Open a card to read the detail.
Why a keyword tool cannot see these questions
Search volume describes what people type into a search box. A buying query put to a chat window is a different register:
Read the detail
- It is longer, phrased as a full sentence.
- It carries the qualifiers a keyword would strip.
- Nobody types their headcount and compliance regime into a search bar. Everybody mentions both when they ask an assistant who to hire.
So the queries worth winning report no volume at all. We hit this on our own site: the planner returned zero for the exact phrasings buyers were using. We published the measurement behind it rather than asserting it.
A tool that reports zero is not telling you the demand is absent. It is telling you the demand moved to a surface it cannot see.
So the query set comes from the answers themselves:
- We ask the engines and read what comes back.
- The queries where a competitor gets named are the demand signal.
- On a first scan we also seed some phrasings from what people are observably searching, so the set is anchored in both places.
What retrieval does with a page
A generative answer is assembled, not recalled. The system runs a search, pulls candidate documents, and writes its response grounded in their text.
Read the detail
Every property that makes a document cheap to use at that step is worth engineering into a draft. The research literature is direct about which ones matter.
The academic work on generative engine optimization measures source-side changes and finds real movement. The authors report gains of up to 40%, and state plainly that the effect varies by domain.
Read the caveat as the finding. What works is a function of the question and the category, which is why a draft aimed at a named question beats a draft aimed at a topic. The engineering falls out of that:
- State the question in a heading, so the match is obvious.
- Answer it in the next sentence, so an extractor gets your position rather than your preamble.
- Carry specifics. A passage with five verifiable facts survives a quotation; a passage of adjectives does not.
- Define terms instead of assuming them. The retrieved chunk arrives without your other pages around it.
What a content agency cannot do: prove the post worked
A good writer can give you a good post. A writer cannot give you the query it was written for, the answer it was written to change, and a reading a week later of whether it did. The Content Engine starts and ends with those:
Read the detail
- Each draft is tied to a buying query your scan recorded as lost.
- It publishes to your site through your own CMS connection: WordPress, Webflow, HubSpot, a GitHub repository or a webhook.
- The next scan re-asks that query on all five engines.
- When an engine starts naming you on that query, the piece is marked cited. That is the result we count.
Traffic and shares are pleasant and beside the point. A post written to be liked delays its answer to hold attention. A post written to be cited states it in the first two sentences, because that is the passage an engine lifts.
A post that leaves its query unmoved is information too. The answer it argued against may lean on a directory, a review site or a rival's comparison page, and the Sources pane shows which. The next draft goes after that source instead of a fourth post on the same topic.
See where you stand first
Run a free scan and find out which of these gaps you actually have.
Get your free auditFrequently asked questions
How are topics chosen?
From your scan, never from a brainstorm. Each draft targets a buying query where an engine named a competitor and skipped you, and the highest-intent gaps go first, so the earliest posts sit closest to a purchase decision.
Will the drafts sound like my company?
They arrive in a clear, specific B2B voice with your positioning and your real service detail in them, and they are yours to edit. Because they lead with substance instead of adjectives, most need a light pass to add your point of view, not a rewrite.
Will AI-drafted content get my site penalized by Google?
Google's spam policies target scaled content abuse, meaning thin mass-produced pages. The definition turns on the page, not on who wrote it. Every draft here is grounded in your positioning and a real buying query, and you review before publishing: the quality floor that keeps a page safe on both Google and the engines.
How long are the posts?
Long enough to answer the question and no longer. Most land between 1,200 and 2,000 words because that is what a thorough answer with checkable specifics takes, but the length follows the question rather than a quota.
I already publish a blog. Do I still need this?
The scan answers that better than we can. If your existing posts already cover the buying queries where engines skip you, your gap is small. Most firms find the opposite, and this closes that specific gap alongside whatever you already publish.
How do I know a post worked?
You watch the query, not the post. The Content Engine tracks each piece against the buying queries it targeted, and the next scan re-asks those queries on all five engines. If the answer starts naming you, the piece is marked cited.
Sources and further reading
- SE Ranking, structured data in AI answers (via Search Engine Land): 71% of pages ChatGPT cites include structured data; 65% for Google AI Mode.
- Aggarwal et al. 2024, Generative Engine Optimization (arXiv:2311.09735): A KDD 2024 paper measuring how source-side changes move visibility in generative engine answers; the authors report gains of up to 40% and caveat that they vary by domain.
- Google Search Essentials, spam policies: Scaled content abuse is the violation, defined by thin mass production rather than by who or what wrote the page.
Related solutions